Atomic Structures and Gram Scale Synthesis of Three Tetrahedral

Article
pubs.acs.org/JACS
Atomic Structures and Gram Scale Synthesis of Three Tetrahedral
Quantum Dots
Alexander N. Beecher,† Xiaohao Yang,‡ Joshua H. Palmer,† Alexandra L. LaGrassa,† Pavol Juhas,§
Simon J. L. Billinge,‡,§ and Jonathan S. Owen*,†
†
Department of Chemistry and ‡Department of Applied Physics and Applied Mathematics, Columbia University, New York, New
York 10027, United States
§
Condensed Matter Physics and Materials Science Department, Brookhaven National Laboratory, Upton, New York 11973, United
States
S Supporting Information
*
ABSTRACT: Luminescent semiconducting quantum dots
(QDs) are central to emerging technologies that range from
tissue imaging to solid-state lighting. However, existing
samples are heterogeneous, which has prevented atomicresolution determination of their structures and obscured the
relationship between their atomic and electronic structures.
Here we report the synthesis, isolation, and structural
characterization of three cadmium selenide QDs with uniform
compositions (Cd 35 Se 20 (X) 30 (L) 30 , Cd 56 Se 35 (X) 42 (L) 42 ,
Cd84Se56(X)56(L)56; X = O2CPh, L = H2N-C4H9). Their
UV-absorption spectra show a lowest energy electronic transition that decreases in energy (3.54 eV, 3.26 eV, 3.04 eV) and
sharpens as the size of the QD increases (fwhm = 207 meV, 145 meV, 115 meV). The photoluminescence spectra of all three
QDs are broad with large Stokes shifts characteristic of trap-luminescence. Using a combination of single-crystal X-ray diffraction
and atomic pair distribution function analysis, we determine the structures of their inorganic cores, revealing a series of pyramidal
nanostuctures with cadmium terminated {111} facets. Theoretical and experimental studies on these materials will open the door
to a deeper fundamental understanding of structure−property relationships in quantum-confined semiconductors.
■
INTRODUCTION
Even the most monodisperse QD samples generally exhibit
dispersity in composition, size, shape, and structure at the
atomic level.8 This heterogeneity results from a growth
mechanism that favors no particular size.1 There are some
examples of so-called “magic-sized” CdSe clusters9−12 whose
preference for certain sizes indicates a high degree of
uniformity, but whose structures remain unknown and, in
some cases, are suspected to be disordered.13,14 Another family
of II-VI materials has been structurally characterized by
SCXRD,15−17 but their unusual stability is made possible by
phenyl chalcogenolate ligands, which produce chalcogen rich
stoichiometries and dominate their optoelectronic properties.18,19 In contrast, most QDs have metal rich stoichiometries
and surfaces bound by carboxylate, phosphonate, and/or amine
ligands.20,21 Here we report the first atomic structures of three
QDs with this characteristic surface motif. This advance brings
us closer to an atomically resolved understanding of the
chemistry and physics of quantum-confined semiconductors.
The unique optoelectronic properties of semiconducting
quantum dots (QDs) are derived from the size and structure
of their inorganic cores as well as a ligand shell that provides
colloidal stability and passivates surface derived electronic
states.1,2 However, a detailed understanding of the relationships
between their atomic structure, optoelectronic properties, and
thermodynamic stability remains elusive because atomic
resolution structures are not known.
Determining the atomic structures of nanomaterials is
notoriously difficult,3 and QDs, in particular, lack exact
structural solutions. Single crystal X-ray diffraction (SCXRD)
has been successfully applied to certain well-defined nanomaterials,4,5 but its applicability to less well-defined materials,
including semiconducting QDs, is limited by the stringent
requirement that diffraction-quality single crystals can be
obtained. More widely available are powder samples of
nanocrystals that are orientationally and spatially disordered.
These samples produce broad, diffuse X-ray scattering features,
which may contain sufficient information to allow structure
determination by total scattering and atomic pair distribution
function (PDF) analysis.6,7 However, to obtain exact atomic
structure solutions, these X-ray techniques require bulk samples
of ordered and atomically precise materials.
© XXXX American Chemical Society
■
RESULTS AND DISCUSSION
To access single-sized, atomically precise QDs, we targeted a
recently discovered class of nanocrystals that exhibit strong size
Received: April 14, 2014
A
dx.doi.org/10.1021/ja503590h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Journal of the American Chemical Society
Article
Figure 1. (a) Formation of CdSe QDs occurs upon combining Cd(O2CR)2, n-alkylamine, and (TMS)2Se precursors in diethyl ether at −78 °C
followed by controlled warming to room temperature. (b−d) In situ monitoring of QD growth with UV−visible absorbance spectroscopy reveals
isosbestic points indicating the direct conversion of smaller QDs into larger ones without the accumulation of intermediates. Spectra were measured
every two minutes.
preference and grow in an unusual “quantized” manner in
which small nanocrystals convert into larger ones bypassing
intermediate sizes. 22 The reported syntheses of these
compounds required elevated reaction temperatures (>60 °C)
to effect precursor conversion, resulting in inseparable mixtures
of QD sizes.12,22,23 We discovered that a highly reactive
selenium precursor, bis(trimethylsilyl)selenide ((TMS)2Se),
converts quantitatively into cadmium selenide upon mixing
with cadmium carboxylates and primary alkylamines at room
temperature or below (Figure 1a). Using 1H nuclear magnetic
resonance (NMR) spectroscopy, we observe the clean
formation of the expected trimethylsilyl ester coproduct
indicating complete and selective (TMS)2Se conversion at
room temperature. The CdSe containing product is conveniently purified by precipitation with acetonitrile and subsequent
filtration, allowing us to selectively prepare a single QD size on
gram scale. A variety of carboxylates derived from sp2-centers
(e.g., substituted benzoates and furoate) and primary aliphatic
amines (e.g., n-alkylamines and 3-phenylpropylamine) can be
substituted for the standard benzoate and n-BuNH2 ligands,
while aliphatic carboxylates and bulkier amines lead to less
stable structures and different growth pathways (Supporting
Information Figure S1).
The high reactivity of (TMS)2Se made it possible to monitor
QD formation at low temperature by in situ UV−visible
absorbance spectroscopy (Figure 1b−d). As an ethereal
solution of n-butylamine (n-BuNH2), cadmium benzoate
(Cd(O2CPh)2), and (TMS)2Se is warmed from −78 °C to
−42 °C (Figure 1b), the precursors react, and the π to π*
transitions of the benzoate ligands become obscured by a much
stronger absorbance from the cadmium selenide product (λmax
= 263 nm). This species subsequently converts into a
previously unobserved species, CdSe(315 nm) (λmax = 315 nm),
at 0 °C. Further warming to room temperature transforms this
QD into a larger one, CdSe(350 nm). Controlled conversion to
larger sizes continues at higher reaction temperatures,
ultimately producing QDs with dimensions of ∼3 nm.12 The
well-defined isosbestic points visible in the spectral data (Figure
1c,d) are characteristic of two families of molecules that
quantitatively interconvert without the formation of strongly
absorbing intermediates. Thus the component spectra are
representative of single QDs. This behavior demonstrates the
reaction’s extraordinary selectivity for specific products that
inhabit local thermodynamic minima and are therefore likely to
have atomically precise structures.
Under these size-selective growth conditions, mixtures of
multiple QDs are avoided, and instead, we prepare and isolate
three single-sized QDs with lowest-energy electronic transitions
(LEETs) at 350, 380, and 408 nm (CdSe(350nm), CdSe(380 nm),
and CdSe(408 nm)). The strategy is remarkable for its simplicity:
gently heating a solution of CdSe(350 nm) completely converts it
into a mixture of CdSe(380 nm) and CdSe(408 nm) that can be
trivially separated by selective precipitation. It should be noted
that CdSe(350nm) slowly grows to the next largest size when
stored in solution at room temperature, particularly if the amine
concentration is lower than 5 mM, suggesting a dynamic
equilibrium with a coordinatively unsaturated species preceding
growth.
The optical properties of these samples resemble those of
conventional QDs with very small dimensions (Figure 2 and
Table 1) exhibiting intense, size-dependent excitonic absorption features. The full width at half-maximum (fwhm) of the
LEET decreases from 207 to 145 meV and then to 115 meV
with increasing QD size, linewidths that are comparable to
those of larger, relatively monodisperse QD ensembles.24 The
observed trend of increasing fwhm with decreasing size is
consistent with predictions that exciton−phonon coupling is
strengthened in small crystallites.25−27 The photoluminescence
(PL) of these samples is broad and substantially Stokes-shifted
but nonetheless size-dependent, characteristics that are typical
of small QDs.12 Notably, larger QDs (LEET ≥ 425 nm)
formed at higher temperatures using similar synthetic methods
display band edge PL in addition to a broad trap-luminescence
feature.12 Thus, the large Stokes shift and broad line width of
B
dx.doi.org/10.1021/ja503590h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Journal of the American Chemical Society
Article
Figure 3. Two views of the Cd35Se20 core structure (Se is orange, Cd
is green). Thermal ellipsoids are set at the 30% probability level.
turned to PDF analysis of total X-ray scattering from QD
powders (see Experimental Section). In brief, the PDF is a
weighted probability distribution of interatomic distances
generated by taking a Fourier transform of the integrated
scattering data. The PDFs of all three QDs are shown in Figure
4. The sharp, well-resolved peaks suggest high symmetry and a
well-defined local structure. The PDF of CdSe(350 nm) (Figure
4a), for example, exhibits sharp features to at least 1.3 nm,
which is consistent with a monodisperse sample of QDs with
dimensions on the order of the single crystal solution. Broad
features visible at higher r originate from interparticle
correlations of a relatively ordered powder.6 As expected
from the optical data, the PDFs of CdSe(380 nm) and
CdSe(408 nm) exhibit sharp features that extend to progressively
longer distances.
While ab initio determination of a structural solution from
the PDF alone is a significant challenge, the SCXRD of
CdSe(350 nm) provided a candidate structural model, a zincblende tetrahedron with the formula Cd35Se20. A simulated
PDF generated from the single crystal structure is in excellent
agreement with the experimental data resulting in very low
residual signal intensity across the fitted range (Rw = 0.14,
fitting range = 1−20 Å). We arrived at similar structural models
for CdSe(380 nm) and CdSe(408 nm) by hypothesizing that their
structures are larger tetrahedra expanded by additional layers of
atoms. Simulated PDFs of pyramids cut from the zinc-blende
lattice closely match the experimental PDFs as shown in Figure
4b,c. The fits are excellent, with lower Rw values of 0.12 for
CdSe(380 nm) and 0.10 for CdSe(408 nm) (fitting range = 1−20
Å). For comparison, previously published simulations of bulk
CdSe (Rw = 0.12, fitting range = 1−40 Å), 2.4 nm CdSe
nanocrystals (Rw = 0.28, fitting range = 1−40 Å), and C60 (Rw =
0.18, fitting range = 1−10 Å) have larger residual signals.14 It is
important to note that it is difficult to compare Rw values
measured across different fitting ranges and for different
materials due to the influence of systematic errors that depend
upon atomic scattering factors and increase with larger fitting
ranges.
To assess the reliability of our solutions, we explored other
candidate structures by simulating their PDFs and comparing
the magnitude of their Rw values. Very different structures,
including those previously reported for bare, stoichiometric
“magic-sized” clusters,9,30−32 are easily ruled out by their
qualitatively distinct PDFs and much larger Rw values
(Supporting Information Figure S3). Simulated PDFs of
known tetrahedral clusters with chalcogenolate surface ligands
and zinc blende-like cores better reproduce the experimental
data (Supporting Information Figure S4);33,34 however, the fit
is still relatively poor, which can perhaps be attributed to the
clusters’ chalcogen rich compositions and distorted zinc blende
Figure 2. Absorbance (298 K, solid black lines), photoluminescence
excitation (PLE, 77 K, dotted red lines) and photoluminescence (77 K,
dashed red lines) spectra of purified QDs. Blue down-pointing arrows
indicate the excitation wavelength for PL and green up-pointing
arrows indicate the emission wavelength used for PLE measurements.
A small concentration of CdSe(380 nm) spontaneously forms from
CdSe(350nm) at room temperature and is visible in the absorption and
PLE spectrum of CdSe(350nm). Similarly, a small concentration of a
larger size is visible in the absorption and PLE spectra of CdSe(408nm).
Table 1. Summary of QD Optical Properties
QD
1st/2nd λ
max.a (nm)
1st/2nd abs.
fwhm (meV)
PL max /Exc.
λb (nm)
PL fwhmb
(meV)
CdSe(350 nm)
CdSe(380 nm)
CdSe(408 nm)
350/333
380/357
408/384
207/304
145/257
115/175
457/346
484/368
529/396
527
471
461
a
To extract the fwhm, absorption spectra were expressed in eV and the
regions corresponding to the lowest energy electronic transition were
fitted to a Gaussian function using least-squares regression analysis.
b
Measured at 77 K.
the PL of CdSe(350nm), CdSe(380 nm), and CdSe(408 nm) appear to
be a consequence of their very small size and are perhaps
related to the strong coupling of the exciton to vibrational
motions, a behavior that is typical of small molecule
fluorophores and bulk materials that undergo exciton selftrapping upon distortion in the excited state.28,29
In order to determine the QDs’ structures, we combined
SCXRD and PDF analysis. Diffraction quality single crystals of
CdSe(350 nm) were grown via diffusion of acetonitrile into a
solution of diethyl ether. These crystals could be redissolved to
verify that the absorption spectrum is unchanged by the
crystallization process. Although disorder prevented modeling
of the organic ligand shell and limited structure refinement to
1.8 Å resolution, the data are sufficient to conclude that
CdSe(350 nm) is a zinc-blende tetrahedron terminated by
cadmium rich {111} facets with the formula: Cd35Se20 (Figure
3 and Table S1 in the Supporting Information).
To obtain structures of CdSe(380 nm) and CdSe(408 nm), which
did not readily form diffraction-quality single crystals, we
C
dx.doi.org/10.1021/ja503590h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Journal of the American Chemical Society
Article
Figure 4. Experimental PDF data (blue) with simulated PDFs (red) overlaid for the three different QDs studied. The quality of the agreement is
very good as shown by low residuals (black) and goodness of fit values, Rw.
structures. A series of models with more subtle differences from
the tetrahedral solutions shown in Figure 4 also tested (Figure
5). Both addition and subtraction of single atoms at the edges,
corners, and faces lead to poorer agreement and increased Rw.
Comparisons of simulated and experimental data were made
across several ranges of r (1−12 Å, 1−16 Å, 1−20 Å, 1−22 Å,
and 1−25 Å) to evaluate any bias caused by inter-QD
correlations at high r or correlations from organic ligands at
low r that are not part of the simulation. For CdSe(380 nm) and
CdSe(408 nm), the results unambiguously support the pristine
tetrahedral structure with purely cadmium enriched surfaces
including four corner Cd atoms. The consistency among
different fitting ranges demonstrates the PDF results are reliable
and not dominated by data collection statistics, for example.
The quality of the fit for CdSe(350 nm) depends on the fitting
range, a result that might be explained by the greater
contribution of inter-QD and organic ligand scattering
correlations in the data. This ambiguity suggests that the
information content in the PDF data alone is marginal for
making this determination. However, the structure solution
with all four corner atoms is preferred for several reasons: (1)
The Rw minima are clustered close to this structure; (2) models
of the larger QDs evaluated over all fitted ranges favor corner
atoms; and (3) removing 1−3 corner atoms would lead to
lower symmetry. These observations strongly support the
tetrahedral assignments shown in Figure 4 for all three sizes.
Due to disorder in the single crystals and the relatively weak
scattering power of the organic ligands, the combined PDF/
SCXRD approach did not allow complete characterization of
the molecular formula. We instead deduce the formulas of
CdSe(350nm),CdSe(380nm), and CdSe(408nm) using inductively
coupled plasma optical emission spectroscopy (ICP-OES),
combustion analysis, infrared absorption, and nuclear magnetic
resonance (NMR) spectroscopies (Table 2, Supporting
Information Figures S5 and S6). The Cd:Se ratios measured
by ICP-OES are consistent with the structures determined by
X-ray diffraction. The Cd:Se ratio fixes the benzoate content
because each cadmium in excess of selenium must be chargebalanced by two benzoate anions to arrive at a neutral QDligand complex. The 1H NMR spectra of each QD shows that
the number of n-butylamine ligands is equal to the number of
benzoate ligands within the 10% uncertainty of the integration.
Thus, in the case of CdSe(350 nm), the 15 excess cadmiums are
balanced by 30 benzoate anions and an additional 30 ± 3 nBuNH2 ligands. This line of reasoning yields the chemical
formulas Cd 35 Se 20 (O 2 CAr)30 (H 2 NR) 30 for CdSe (350 nm) ,
Cd 5 6 Se 3 5 (O 2 CAr) 4 2 (H 2 NR) 4 2 for CdSe ( 3 8 0 n m ) , and
Figure 5. PDF structure modeling. (a) Atomic structure models of
CdSe(350 nm) and their agreement factor (Rw) to the experimental data
across the range r = 1−20 Å. The Cd and Se atoms are colored in
magenta and green, respectively. (b) Change in Rw for CdSe(350 nm)
(bottom), CdSe(380 nm) (middle), and CdSe(408 nm) (top) as a function
of single atom additions/subtractions to the pristine tetrahedron. Five
different fitting ranges were evaluated: 1−12 Å (red dots), 1−16 Å
(green up-pointing triangles), 1−20 Å (blue squares), 1−22 Å (black
down-pointing triangles), and 1−25 Å (turquoise pentagons). ΔRw is
calculated relative to the best fitting model, which is generally the
pristine tetrahedron (ΔRw = Rw − Rw(min)).
D
dx.doi.org/10.1021/ja503590h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Journal of the American Chemical Society
Article
Table 2. Summary of QD Structural Properties
QD
chemical formula
edge length
(nm)
coord.
sites
# of
ligandsa
ICP-OES.
Cd:Se ratiob
theor. Cd:Se
ratio
measured
% mass C
theor.
% mass C
CdSe(350 nm)
CdSe(380 nm)
CdSe(408 nm)
Cd35Se20(O2CPh)30(H2NR)30
Cd56Se35(O2CPh)42(H2NR)42
Cd84Se56(O2CPh)56(H2NR)56
1.71
2.14
2.57
60
84
112
60
84
112
1.71 ± 0.14
1.61 ± 0.17
1.53 ± 0.06
1.75
1.60
1.50
32.92
31.63
27.96
34.90
32.19
29.87
a
Uncertainty in individual measurements of amine content is 10% due to the difficulty of accurately integrating the broad NMR signals. Repeated
measurements support a 1:1 amine to benzoate ratio for each QD (CdSe(350 nm) average ratio 1.045 with a standard deviation of 0.050; CdSe(380 nm)
average ratio 1.025 with a standard deviation of 0.066). bDeterminations of cadmium and selenium content were made by inductively coupled
plasma optical emission spectroscopy (ICP-OES). CdSe(350 nm) was measured a total of five times. CdSe(380 nm) and CdSe(408 nm) were each
measured three times.
Cd84Se56(O2CAr)56(H2NR)56 for CdSe(408 nm). However, a
range of ligand compositions is possible and cannot be ruled
out given the scatter in the elemental analysis data.
The number of ligands found on each QD matches the
number of vacant coordination sites assuming four-coordinate
cadmium centers and monodentate benzoate coordination. On
a cadmium selenide {111} facet, the density of coordination
sites (6.2 sites/nm2) is higher than the packing density of
crystalline alkane chains (4.9 chains/nm2) and the binding of
one ligand to each coordination site over large areas is
unlikely.35,36 On these small, tetrahedral QDs, however, the
effective curvature increases the volume available to the ligand
shell by 2−3 times relative to a planar surface of equivalent area
(Supporting Information Table S2). Using our measured
formulas, we estimate that the density of ligand atoms (C, H,
N, O) increases from 80 atoms/nm3 to 90 atoms/nm3 as the
tetrahedron increases in size and the effective curvature
decreases. These densities are lower than what is found in
typical crystalline alkylammonium benzoate salts (∼105 atoms/
nm3) indicating that the proposed chemical formulas are not
limited by the packing density of ligands despite their high
coverage (see Supporting Information for additional discussion). One ligand on every coordination site leads to a surface
without dangling bonds, perhaps explaining the special stability
of these structures. This high ligand density further supports
the hypothesis that the photoexcited QD reorganizes according
to the Franck−Condon principle rather than radiating from a
distribution of luminescent midgap excitonic states.28,29
Knowledge of each QD’s structure allows us to precisely
analyze the size-dependence of the band gap at these unusually
small sizes. By plotting an effective radius against the energy of
the LEET for both our QDs and previously reported QDs,37 we
observe a nearly identical size dependence in the two sample
sets (Figure 6). We fit the combined data sets to a fourth-order
polynomial to generate a sizing curve that is nearly identical to
the one determined previously,37 but better supported at
energies greater than 3 eV.
The atomically precise compositions of CdSe(350 nm),
CdSe(380 nm), and CdSe(408 nm) made their structural solution
possible. In a few cases, the formulas and calculated structures
of other cadmium selenide species that adopt preferred sizes
have been assigned on the basis of laser-desorption ionization
mass spectrometry (LDI-MS) measurements.9,32,38 Among the
range of species observed in these spectra, (CdSe)13, (CdSe)19,
(CdSe)33, and (CdSe)34 are found in greatest abundance.9
However, this technique produces the same fragments from
much larger nanocrystals and bulk CdSe.9,22,23 Indeed, LDI-MS
studies of CdSe(350 nm) also produce signals from (CdSe)13,
(CdSe)19, (CdSe)34, and other fragments (see Supporting
Information Figure S7 and S8). The lack of organic ligands
suggests that these fragments result from the special stability of
Figure 6. Our QDs (red circles) exhibit very similar size dependence
to the experimental (solid squares) and calculated (open squares)
results used by Jasieniak et al. to create a sizing curve. Merging our two
data sets yields a new empirical sizing relationship d(nm) = 49.9019 −
(0.470699)λ + (1.66480 × 10−3)λ2 − (2.57705 × 10−6)λ3 + (1.49956
× 10−9)λ4 (green, solid line) that is similar to the relationship found by
Jasieniak et al. (black, dashed line), but better supported in the small
size regime.37
gas-phase clusters formed by laser ablation rather than ones
stabilized in solution. In contrast, the stability of CdSe(350 nm),
CdSe(380 nm), and CdSe(408 nm) is a consequence of their surface
termination and the high ligand coverage; at these sizes, a
tetrahedron is uniquely suited to minimize the number of
unfilled coordination sites per surface atom while simultaneously maintaining a low surface-area-to-volume ratio.
■
CONCLUSION
The reported structures and formulas provide a direct view into
the origins of QD stability and optical properties without
ambiguities from distributions of crystal sizes and shapes
and/or electronic transitions associated with the ligand shell.
The dynamic nature of these materials is remarkable; complex,
pyramidal structures form rapidly at low temperatures and
readily convert into larger QDs without accumulation of
intermediate structures. In addition, we demonstrate the power
of PDF methods for determining the structure of homogeneous
nanomaterials. This approach holds great promise for atomic
structure solutions generally, and particularly as ab initio
structure-searching methods develop. Studies of this kind will
enable detailed experimental and first-principle investigations
into the relationship between geometric and electronic
structure on the nanoscale.
E
dx.doi.org/10.1021/ja503590h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Journal of the American Chemical Society
■
Article
carboxylic acid and dimethylcadmium following a procedure adapted
from Hendricks et al.39 Briefly, in a nitrogen filled glovebox and in the
dark, 1.98 equiv of the carboxylic acid are dissolved in tetrahydrofuran
to which 1 equiv of dimethylcadmium is added dropwise. Vigorous
bubbling occurs and the solution is stirred for 30 min. The solvent is
distilled off under vacuum leaving a white powder, which is then dried
overnight under vacuum at 100 °C.
Synthesis and Isolation of bis(trimethylsilyl)selenide (Se(Si(CH3)3)2). The following two-step procedure is a variation on the one
pot method reported by Detty and Seidler that proved more reliable
and higher yielding.40
Synthesis and Isolation of Lithium Selenide (Li2Se). In a
Teflon sealed Schlenk tube under argon, lithium triethylborohydride in
tetrahydrofuran (100 mL, 1 M) is chilled to 0 °C, and selenium pellets
(3.91 g, 49.5 mmol) are added. The mixture is allowed to warm to
room temperature and stirred for an hour resulting in a cloudy, white
suspension. Tetrahydrofuran and triethylborane are removed by
distillation under vacuum (CAUTION: Triethylborane is extremely
f lammable and should be carefully quenched prior to disposal), and the
resulting white solid is triturated with pentane (50 mL) and collected
on a fine glass frit in the glovebox. The white powder is then washed
with toluene (50 mL) followed by diethyl ether (50 mL) and then
dried under reduced pressure at 150 °C taking care to avoid losing the
very fine white powder upon evacuation. (4.28 g; 46 mmol, 93.4%
yield). It is important to carefully protect lithium selenide from oxygen
to obtain a pure, colorless product. A known quantity of the powder
and an internal standard (N,N,N′,N′-tetramethylethylenediamine) are
dissolved in methanol-d4 to check for the presence of organic
impurities by 1H NMR spectroscopy; typical syntheses produced a
product that is >98% free from tetrahydrofuran or ethyl containing
impurities.
Synthesis and Isolation of bis(trimethylsilyl)selenide (Se(Si(CH3)3)2). To a nitrogen filled Schlenk tube containing lithium
selenide (4.50 g, 48.5 mmol), bromotrimethylsilane (14.47 g, 94.52
mmol) is added via syringe. The suspension is degassed using the
freeze−pump−thaw technique, sealed under vacuum, and heated to
100 °C for 24 h. The flask is then attached to a bulb-to-bulb, vacuum
transfer apparatus and a clear, colorless oil is distilled from the
suspension with the help of a heat gun (8.54 g; 37.9 mmol; 80.2%
yield). If necessary, residual bromotrimethylsilane or hexamethyldisiloxane can be removed by partial vacuum distillation (45 °C, 20 Torr)
leaving pure Se(Si(CH3)3)2. In the event that accidental contamination
of the reaction mixture or the starting material with oxygen occurs, an
impure yellow product is obtained upon distillation. This material
should be purified by addition of tri-n-octylphosphine and
redistillation prior to long-term storage. 1H NMR (C6D6, 400
MHz): δ = 0.38. 13C{1H} NMR (125.8 MHz, C6D6): δ = 4.65
(-CH3) ppm.
Synthesis and Isolation of CdSe(350 nm). Cadmium benzoate
(2.00 g, 5.43 mmol) and n-butylamine (0.993 g, 13.58 mmol) are
dissolved in diethyl ether (20 mL), and the solution is cooled to 0 °C
in an ice water bath. In a separate flask, a solution of
bis(trimethylsilyl)selenide (0.612 g, 2.72 mmol) and diethyl ether
(20 mL) is cooled to 0 °C and then cannula-transferred into the
cadmium precursor solution. The mixture is allowed to warm to room
temperature. After stirring for 30 min, the reaction mixture is
concentrated under reduced pressure, and the flask is brought into the
glovebox. The sticky white solid is triturated with acetonitrile (10 mL)
and the powder is collected by centrifugation (3 min at 7000 rpm).
After decanting the supernatant, the pellet is dissolved in a minimal
amount of diethyl ether (<5 mL), reprecipitated with acetonitrile (10
mL), and again collected by centrifugation. This process is repeated
once more. The purified product is dried from ether under vacuum
overnight at room temperature to obtain 1.02 g of fine white powder.
1
H NMR (500 MHz, C6D6): δ = 0.60 (br, 3H, Me), 0.99 (br, 2H,
-CH2-), 1.36 (br, 2H, -CH2-), 2.98 (br, 2H, -NH2), 4.18 (br, 2H,
-CH2-), 7.06 (br, 2H, Ar-H), 7.28 (br d, 1H, Ar-H), 8.15 (br, 2H, ArH) ppm. 13C{1H} NMR (125.8 MHz, C6D6): δ = 13.88 (-CH3), 20.19
(-CH2-), 34.85 (-CH2-), 42.78 (-CH2-), 130.85 (Ar), 131.11 (Ar),
EXPERIMENTAL SECTION
General Considerations. Benzoic acid (≥99.5%), benzoic
anhydride, dichloromethane (≥99.5%), selenium pellets (≥99.99%),
Superhydride solution (1.0 M lithium triethylborohydride in
tetrahydrofuran), and tetrahydrofuran (≥99.0%) were purchased
from Sigma Aldrich and used as received. Benzene-d6 (99.6%) and
anhydrous acetonitrile (99.5%) were purchased from Sigma Aldrich,
shaken with activated alumina, filtered, and stored over 4 Å molecular
sieves for at least 24 h prior to use. Diethyl ether, pentane, and toluene
were dried over alumina columns and stored over 4 Å molecular sieves
for at least 24 h prior to use. Bromotrimethylsilane (97%), nbutylamine (99%), n-octylamine, and tetramethylethylenediamine
(TMEDA) were purchased from Sigma Aldrich and dried over
calcium hydride, distilled, and stored in a nitrogen glovebox.
Methanol-d4 was purchased from Cambridge Isotope Laboratories
and used as received. Dimethylcadmium was purchased from Strem,
vacuum distilled before use, and stored in a nitrogen atmosphere
glovebox. CAUTION: Dimethylcadmium is an extremely toxic liquid and
due to its volatility and air-sensitivity should only be handled by a highly
trained and skilled scientist.
Unless otherwise indicated, all manipulations were performed under
air-free conditions using standard Schlenk techniques or a nitrogen
atmosphere glovebox. NMR spectra were recorded on Bruker Avance
III 400 and 500 MHz instruments and internally referenced to the
resonances of protio-impurities in the deuterated solvent. 1H NMR
spectra were acquired with sufficient delay to allow complete
relaxation between pulses (15 s). Coupling constants are reported in
hertz. UV−visible absorption data were obtained using a Perkin-Elmer
Lambda 650 spectrophotometer equipped with deuterium and
tungsten halogen lamps. In situ absorbance measurements were
performed using a fiber optic dip probe. Photoluminescence and
photoluminescence excitation spectra were recorded using a
FluoroMax-4 from Horiba Scientific. Diffuse Reflectance Infrared
Fourier Transform Spectroscopy (DRIFTS) was performed on a
Nicolet 6700 FT-IR from Thermo Fisher equipped with a Harrick
Praying Mantis Diffuse Reflection Accessory.
Synthesis and Isolation of Cadmium Benzoate ((Cd(C7H5O2)2)3(CH3CN)1). To a Schlenk flask, cadmium oxide (18.11 g,
141 mmol), benzoic acid (103.31 g, 846 mmol), and benzoic
anhydride (38.23 g, 169 mmol) are added, and the flask is sealed under
vacuum. The red-brown mixture is heated with stirring to 180 °C to
achieve a colorless melt, which resolidifies as the mixture is cooled to
room temperature. Excess benzoic acid and benzoic anhydride are
removed by washing the solid on a fritted funnel with toluene (3 ×
200 mL toluene) and then dichloromethane (2 × 200 mL
dichloromethane). The resulting powder is dissolved in a mixture of
100 mL tetrahydrofuran and 200 mL acetonitrile and concentrated by
distillation with heat causing a milky white solid to precipitate. The
cooled suspension is filtered and the powder dried under vacuum at
100 °C overnight to yield 24.75 g (47.7%) of free flowing white
powder. The product is insoluble or sparingly soluble in most
common organic solvents with the exceptions of coordinating solvents
such as tetrahydrofuran, 1,4-dioxane, and benzonitrile. Single crystals
suitable for X-ray crystallography were grown upon cooling a saturated
acetonitrile/tetrahydrofuran (∼95:5) solution (see Figure S9 and
Table S4 in the Supporting Information). Solutions in benzene-d6
were prepared for analysis by 1H NMR spectroscopy by adding 2 equiv
of n-butylamine. 1H NMR (400 MHz, C6D6): δ = 0.66 (t, 3JH‑H = 7,
6H, Me), 0.73 (s, ACN, -CH3), 1.00 (m, 4H, -CH2-), 1.18 (m, 4H,
-CH2-), 2.61 (s, 4H, -NH2), 2.67 (t, 3JH‑H = 7, 4H, -CH2-), 7.21 (t,
3
JH‑H = 7, 2H, Ar-H), 7.28 (t, 3JH‑H = 7, 4H, Ar-H), 8.55 (d, 3JH‑H = 8,
4H, Ar-H) ppm. 13C{1H} NMR (125.8 MHz, C6D6): δ = 0.29 (ACN,
-CH3), 13.92 (-CH3), 20.15 (-CH2-), 35.28 (-CH2-), 42.53 (-CH2-),
128.21 (m-Ar), 130.63 (o-Ar), 131.05 (p-Ar), 136.44 (1-Ar), 174.31
(-CO2-) ppm. Anal. Calcd for Cd3C44H33O12N: C, 47.83; H, 3.01; N,
1.27. Found: C, 47.77; H, 3.08; N, 1.18.
Synthesis and Isolation of Other Cadmium Carboxylates.
Cadmium bis(4-thiomethylbenzoate), cadmium bis(4-bromobenzoate), and cadmium bis(furoate) are prepared from the desired
F
dx.doi.org/10.1021/ja503590h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Journal of the American Chemical Society
Article
Core Structure Determination. Least-squares refinement led
initially to the Cd31Se20 core solution shown in Supporting
Information Figure S2a, which has an R1 of 0.1969 after SQUEEZE
is applied. Additional modeling of residual electron density in the
Cd31Se20 core prior to the application of SQUEEZE leads to the
Cd35Se20 core solution (Supporting Information Figure S2b), which
has a lower R1 of 0.1822 after SQUEEZE is applied. Anisotropic
refinement with the ISOR restraint was applied to all atoms.
Additional Q peaks could be observed near the surface of the
Cd35Se20 core structure, but attempts to model this density were
unsuccessful.
Modeling Organic Moieties. It was not possible to model organic
moieties in the unit cell. This has been observed previously in metal
chalcogenide single crystals and has been attributed to gross
disorder.46,47 The SQUEEZE procedure in PLATON finds large
void spaces in the unit cell filled with electron density. For the
Cd31Se20 core, SQUEEZE calculates a residual void volume of 52,746
Å3, containing 21,766 electrons; for the Cd35Se20 core, SQUEEZE
calculates a residual void volume of 52,461 Å3, containing 20,732
electrons.
Powder X-ray Diffraction and Pair Distribution Function
Analysis. X-ray powder diffraction experiments were performed on
the X17A beamline at the National Synchrotron Light Source (NSLS)
at Brookhaven National Laboratory. Diffraction data were collected
using the rapid acquisition pair distribution function (RA-PDF)
technique at 100 K with an X-ray energy of 38.872 keV.48 A PerkinElmer flat-panel 2D detector was used for data collection with a
sample distance of 118.269 mm. The diffraction pattern of a Ni metal
standard was measured to calibrate detector geometry, and the
scattering signal of an empty Kapton tube was measured and
subtracted as the background. The 2D diffraction pattern was
integrated to give 1D diffraction intensity in q-space using the
SrXplanar program 49 and transformed into the PDF using
PDFgetX3.50
The atomic PDF G(r) describes the probability of finding atom
pairs at a certain distance, r. Experimentally, the PDF is a sine Fourier
transformation of powder diffraction data using
135.63 (Ar), 174.54 (-CO2-) ppm. Anal.: act. 32.92, 4.55, 3.43; theo.
34.90, 4.23, 3.70.
Synthesis and Isolation of CdSe(380 nm) and CdSe(408 nm).
Under argon, CdSe(350 nm) (500 mg) is dissolved in toluene (35 mL)
producing a colorless solution. After heating at 85 °C for 30 min, a
yellow precipitate forms. The suspension is brought into the glovebox
and centrifuged, separating the yellow CdSe(408 nm) precipitate by
centrifugation. The supernatant containing CdSe(380 nm) is decanted
and the yellow pellet is washed with toluene to remove remaining
CdSe(380 nm). The supernatant is concentrated to dryness under
vacuum, the residue triturated with acetonitrile (10 mL), and the solids
collected by centrifugation. Both samples are dried under vacuum at
room temperature yielding 264 mg of CdSe(380 nm) and 174 mg of
CdSe(408 nm). Solutions of CdSe(408 nm) in benzene-d6 were prepared
for analysis by 1H NMR spectroscopy by performing a ligand exchange
to replace native n-butylamine ligands with n-octylamine ligands. In
brief, the ligand exchange is accomplished by dissolution of
CdSe(408 nm) in a 10 mM solution of n-octylamine in toluene, followed
by precipitation with acetonitrile, and centrifugation. The pellet is
submitted to this process twice more before drying the product under
vacuum. 1H NMR CdSe(380 nm) (500 MHz, C6D6): δ 0.64 (br, 3H,
Me), 1.05 (br, 2H, -CH2-), 1.42 (br, 2H, -CH2-), 3.02 (br, 2H, -NH2),
3.94 (br, 2H, -CH2-), 7.02 (br, 2H, Ar-H), 7.11 (br, 1H, Ar-H), 8.08
(br, 2H, Ar-H) ppm. 1H NMR CdSe(408 nm) (400 MHz, C6D6): δ 0.91
(br, 3H, Me), 1.11 (br d, 10H, -CH2-), 1.26 (br, 2H, -CH2-), 2.99 (br,
2H, -NH2), 4.08 (br, 2H, -CH2-), 7.15 (br, 3H, Ar-H), 8.15 (br, 2H,
Ar-H) ppm. C,H,N analysis of CdSe(380 nm): act. 31.63, 3.99, 2.88;
theo. 32.19, 3.90, 3.42. C,H,N analysis of CdSe(408 nm): act. 27.96, 3.30,
2.37; theo. 29.87, 3.62, 3.17.
In Situ UV−vis Absorption. In situ UV−visible absorption data
(Figure 1b−d) were collected using a Perkin-Elmer Lambda 650 UV−
vis spectrometer equipped with a Mono-Fiber Transfer-Optic that
redirects the beam into a fiber optic dip probe. Under nitrogen, a 100
mL three-neck flask fitted with a dip probe adapter, argon inlet, and a
thermocouple adapter is charged with an ethereal solution (60 mL) of
Cd(O2CPh)2 (2.18 mM) and n-BuNH2 (5.46 mM). This solution is
chilled to −78 °C and a solution of TMS2Se in ether (15 mL, 4.37
mM) is added dropwise with vigorous stirring. The temperature is
raised to −42 °C by submerging the flask in an acetonitrile dry ice
bath. Spectra were recorded from 250 to 500 nm every two minutes.
After the growth of a signal at 263 nm reaches completion, the
temperature is raised to 0 °C by submerging the flask in an ice water
bath. After the growth of signals at 315 and 300 nm reach completion,
the temperature is raised to 22 °C.
Empirical Sizing Formula Determination. Data for our
quantum dots (QDs) were combined with data from Jasieniak et al.
and then fit to a fourth-order polynomial (diameter in nanometers as a
function of position of the lowest energy electronic transition in units
of nanometers) using least-squares regression analysis.37 We modeled
the volume of our QDs by summing the volumes of all Cd2+ and Se2−
ions (ionic radii of 109 and 184 pm respectively).41 These volumes
were used to calculate the radius of a sphere with an equivalent
volume.
Single Crystal X-ray Diffraction. Collection and Refinement
Strategy. Single crystal X-ray diffraction (SCXRD) data were collected
on a Bruker Apex II diffractometer, and crystal data, data collection,
and refinement parameters are summarized in Supporting Information
Table S1. The structures were solved using direct methods and
standard difference map techniques, and were refined by full-matrix
least-squares procedures on F2 with SHELXTL (Version 2008/4).42,43
All atoms not found and refined were treated as a diffuse contribution
to the electron density without specific atomic positions using
SQUEEZE/PLATON.44,45
Unit Cell and Space Group Determination. Data were first
integrated and processed using the primitive reduced cell found in
APEX2, and the initial solution was transformed to the space group
Cmcm using PLATON. At this point, the data were reintegrated using
the C-centered orthorhombic unit cell reported in Supporting
Information Table S1, and structure refinement was performed.
G(r ) =
2
π
∫Q
Q max
Q [S(Q ) − 1]sin Qr dQ
min
where Q is the magnitude of the scattering vector, and S(Q) is the total
scattering function. S(Q) is generated by normalizing and correcting
the coherent scattering intensity Ic(Q) using
S(Q ) =
Ic(Q ) − ⟨f 2 ⟩
⟨f ⟩2
+1
where f is the X-ray scattering factor averaged over all elements in the
sample. Qmin = 0.8 Å−1 and Qmax x = 22.5 Å−1 are limits placed on the Q
range in the PDF transformation. The minimum value is determined
by the beamstop, and the maximum value is chosen to reduce noise
originating in the high-Q region.
Theoretically, the PDF can be calculated from an atomic structural
model:
G(r ) =
1
Nr
∑
i≠j
fi f j
⟨f ⟩2
δ(r − rij)
where the Dirac delta function is expanded to a Gaussian-like function
after including thermal parameters.51 We used SrFit to optimize the
parameter set giving a structural model whose PDF is most consistent
with the experimental PDF. The quality of the fit is characterized by
the residual function
Rw =
N
∑i = 1 [Gobs(ri) − Gcalc (ri ; P⃗ )]2
N
2
∑i = 1 Gobs
(ri)
where Gobs is the experimental PDF, Gcalc is the calculated PDF, and P⃗
is the set of refinable parameters used in the structure model.
G
dx.doi.org/10.1021/ja503590h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Journal of the American Chemical Society
Article
The atomic structures of the QDs reported here are cut from the
CdSe zincblende lattice to a tetrahedral shape with optional addition
or removal of atoms from the corners or edges. Identical isotropic
atomic displacement parameters (Uiso) are assigned for all atoms of a
particular element in a given structure. To optimize fits, we refined Uiso
and the lattice parameters of the base zincblende lattice.
■
(8) Chen, O.; Zhao, J.; Chauhan, V. P.; Cui, J.; Wong, C.; Harris, D.
K.; Wei, H.; Han, H.-S.; Fukumura, D.; Jain, R. K.; Bawendi, M. G.
Nat. Mater. 2013, 12, 445.
(9) Kasuya, A.; Sivamohan, R.; Barnakov, Y. A.; Dmitruk, I. M.;
Nirasawa, T.; Romanyuk, V. R.; Kumar, V.; Mamykin, S. V.; Tohji, K.;
Jeyadevan, B.; Shinoda, K.; Kudo, T.; Terasaki, O.; Liu, Z.; Belosludov,
R. V.; Sundararajan, V.; Kawazoe, Y. Nat. Mater. 2004, 3, 99.
(10) El-Sayed, M. A. Acc. Chem. Res. 2004, 37, 326.
(11) Bowers, M. J.; McBride, J. R.; Rosenthal, S. J. J. Am. Chem. Soc.
2005, 127, 15378.
(12) Cossairt, B. M.; Owen, J. S. Chem. Mater. 2011, 23, 3114.
(13) Cossairt, B. M.; Juhas, P.; Billinge, S. J. L.; Owen, J. S. J. Phys.
Chem. Lett. 2011, 2, 3075.
(14) Yang, X.; Masadeh, A. S.; McBride, J. R.; Božin, E. S.; Rosenthal,
S. J.; Billinge, S. J. L. Phys. Chem. Chem. Phys. 2013, 15, 8480.
(15) Dance, I. G.; Choy, A.; Scudder, M. L. J. Am. Chem. Soc. 1984,
106, 6285.
(16) Herron, N.; Calabrese, J. C.; Farneth, W. E.; Wang, Y. Science
1993, 259, 1426.
(17) Vossmeyer, T.; Reck, G.; Katsikas, L.; Haupt, E. T. K.; Schulz,
B.; Weller, H. Science 1995, 267, 1476.
(18) Soloviev, V. N.; Eichhöfer, A.; Fenske, D.; Banin, U. J. Am.
Chem. Soc. 2000, 122, 2673.
(19) Soloviev, V. N.; Eichhöfer, A.; Fenske, D.; Banin, U. J. Am.
Chem. Soc. 2001, 123, 2354.
(20) Owen, J. S.; Park, J.; Trudeau, P. E.; Alivisatos, A. P. J. Am.
Chem. Soc. 2008, 130, 12279.
(21) Fritzinger, B.; Capek, R. K.; Lambert, K.; Martins, J. C.; Hens, Z.
J. Am. Chem. Soc. 2010, 132, 10195.
(22) Kudera, S.; Zanella, M.; Giannini, C.; Rizzo, A.; Li, Y.; Gigli, G.;
Cingolani, R.; Ciccarella, G.; Spahl, W.; Parak, W. J.; Manna, L. Adv.
Mater. 2007, 19, 548.
(23) Evans, C. M.; Love, A. M.; Weiss, E. A. J. Am. Chem. Soc. 2012,
134, 17298.
(24) Norris, D. J.; Bawendi, M. G. Phys. Rev. B 1996, 53, 16338.
(25) Klein, M. C.; Hache, F.; Ricard, D.; Flytzanis, C. Phys. Rev. B
1990, 42, 11123.
(26) Efros, A. L.; Ekimov, A. I.; Kozlowski, F.; Petrova-Koch, V.;
Schmidbaur, H.; Shumilov, S. Solid State Commun. 1991, 78, 853.
(27) Scamarcio, G.; Spagnolo, V.; Ventruti, G.; Lugará, M.; Righini,
G. C. Phys. Rev. B 1996, 53, R10489.
(28) Turro, N. J.; Ramamurthy, V.; Scaiano, J. C. Modern molecular
photochemistry of organic molecules; University Science Books:
Sausalito, Calif., 2010.
(29) Song, K. S.; Williams, R. T. Self-Trapped Excitons; Springer
Series in Solid-State Sciences; Springer-Verlag: New York, 1993; Vol.
105.
(30) Nguyen, K. A.; Day, P. N.; Pachter, R. J. Phys. Chem. C 2010,
114, 16197.
(31) Nguyen, K. A.; Pachter, R.; Day, P. N. J. Chem. Theory Comput.
2013, 9, 3581.
(32) Wang, Y.; Liu, Y.-H.; Zhang, Y.; Wang, F.; Kowalski, P. J.;
Rohrs, H. W.; Loomis, R. A.; Gross, M. L.; Buhro, W. E. Angew. Chem.,
Int. Ed. 2012, 51, 6154.
(33) Vossmeyer, T.; Reck, G.; Schulz, B.; Katsikas, L.; Weller, H. J.
Am. Chem. Soc. 1995, 117, 12881.
(34) Eichhöfer, A. Eur. J. Inorg. Chem. 2005, 2005, 1245.
(35) Anderson, N. C.; Hendricks, M. P.; Choi, J. J.; Owen, J. S. J. Am.
Chem. Soc. 2013, 135, 18536.
(36) Lüth, H.; Nyburg, S. C.; Robinson, P. M.; Scott, H. G. Mol.
Cryst. Liq. Cryst. 1974, 27, 337.
(37) Jasieniak, J.; Smith, L.; van Embden, J.; Mulvaney, P.; Califano,
M. J. Phys. Chem. C 2009, 113, 19468.
(38) Wang, F.; Richards, V. N.; Shields, S. P.; Buhro, W. E. Chem.
Mater. 2014, 26, 5.
(39) Hendricks, M. P.; Cossairt, B. M.; Owen, J. S. ACS Nano 2012,
6, 10054.
(40) Detty, M. R.; Seidler, M. D. J. Org. Chem. 1982, 47, 1354.
(41) Shannon, R. D. Acta Crystallogr., Sect. A 1976, 32, 751.
ASSOCIATED CONTENT
* Supporting Information
S
UV−vis measurements of QD derivatives, stability, and
stoichiometry, SCXRD solutions, crystal refinement data,
PDF models of previously reported CdSe materials, 1H
NMR, and DRIFTS measurements, a discussion of ligand
packing on a tetrahedron, and MALDI-MS data. This material
is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected].
Author Contributions
Alexander N. Beecher and Xiaohao Yang contributed equally.
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS
We thank Dr. Joshua Choi for assistance with PDF measurements, Ava Kreider-Mueller and Prof. Gerard Parkin for
assistance with SCXRD, Paul Kowalski of Bruker for assistance
with LDI-MS analysis, Zachariah Norman for helpful
discussions, Ruth Pachter for sharing coordinates of computed
magic size clusters, and Paul Mulvaney and Jacek Jasieniak for
sharing size-dependent optical data. Work on QD synthesis and
growth kinetics was supported by the National Science
Foundation through contract number NSF-CHE-1151172.
PDF measurements were supported by the Center for ReDefining Photovoltaic Efficiency Through Molecule Scale
Control, an Energy Frontier Research Center funded by the
U.S. Department of Energy, Office of Science, Office of Basic
Energy Sciences under Award Number DE-SC0001085. PDF
simulations and structure reliability studies were funded by
Laboratory Directed Research and Development (LDRD)
Program 12-007 (Complex Modeling) at the Brookhaven
National Laboratory. X-ray experiments were carried out at the
National Synchrotron Light Source, Brookhaven National
Laboratory, which is supported by the U.S. Department of
Energy, Division of Materials Sciences and Division of
Chemical Sciences, DE-AC02-98CH10886. We thank the
National Science Foundation (CHE-0619638) for the acquisition of an X-ray diffractometer.
■
REFERENCES
(1) Murray, C. B.; Norris, D. J.; Bawendi, M. G. J. Am. Chem. Soc.
1993, 115, 8706.
(2) Alivisatos, A. P. Science 1996, 271, 933.
(3) Billinge, S. J. L.; Levin, I. Science 2007, 316, 561.
(4) Jadzinsky, P. D.; Calero, G.; Ackerson, C. J.; Bushnell, D. A.;
Kornberg, R. D. Science 2007, 318, 430.
(5) Desireddy, A.; Conn, B. E.; Guo, J.; Yoon, B.; Barnett, R. N.;
Monahan, B. M.; Kirschbaum, K.; Griffith, W. P.; Whetten, R. L.;
Landman, U.; Bigioni, T. P. Nature 2013, 501, 399.
(6) Juhás, P.; Cherba, D. M.; Duxbury, P. M.; Punch, W. F.; Billinge,
S. J. L. Nature 2006, 440, 655.
(7) Young, C. A.; Goodwin, A. L. J. Mater. Chem. 2011, 21, 6464.
H
dx.doi.org/10.1021/ja503590h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Journal of the American Chemical Society
Article
(42) Sheldrick, G. M. SHELXTL, An Integrated System for Solving,
Refining and Displaying Crystal Structures from Diffraction Data;
University of Göttingen, Göttingen, Federal Republic of Germany,
1981.
(43) Sheldrick, G. M. Acta Crystallogr., Sect. A 2008, 64, 112.
(44) Spek, A. L. PLATON, A Multipurpose Crystallographic Tool,
Utrecht University, Utrecht, The Netherlands, 2005.
(45) Spek, A. L. J. Appl. Crystallogr. 2003, 36, 7.
(46) Wu, T.; Zhang, Q.; Hou, Y.; Wang, L.; Mao, C.; Zheng, S.-T.;
Bu, X.; Feng, P. J. Am. Chem. Soc. 2013, 135, 10250.
(47) Zheng, N.; Bu, X.; Wang, B.; Feng, P. Science 2002, 298, 2366.
(48) Chupas, P. J.; Qiu, X.; Hanson, J. C.; Lee, P. L.; Grey, C. P.;
Billinge, S. J. L. J. Appl. Crystallogr. 2003, 36, 1342.
(49) Yang, X.; Juhas, P.; Billinge, S. J. L. arXiv:1309.3614, 2013.
(50) Juhás, P.; Davis, T.; Farrow, C. L.; Billinge, S. J. L. J. Appl.
Crystallogr. 2013, 46, 560.
(51) Egami, T.; Billinge, S. J. L. Underneath the Bragg Peaks:
Structural Analysis of Complex Materials; Pergamon Materials Series,
Second.; Elsevier: Amsterdam, 2012; Vol 16.
I
dx.doi.org/10.1021/ja503590h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX